Neuron

Review

Prediction as a Humanitarian and PragmaticContribution from Human Cognitive Neuroscience

John D.E. Gabrieli,1,2,3,* Satrajit S. Ghosh,1,4 and Susan Whitfield-Gabrieli11Poitras Center for Affective Disorders Research at the McGovern Institute for Brain Research2Department of Brain and Cognitive Sciences3Institute for Medical Engineering and ScienceMassachusetts Institute of Technology, Cambridge, MA 02139, USA4Department of Otology and Laryngology, Harvard Medical School, Boston, MA 02115, USA*Correspondence: gabrieli@mit.eduhttp://dx.doi.org/10.1016/j.neuron.2014.10.047

Neuroimaging has greatly enhanced the cognitive neuroscience understanding of the human brain and itsvariation across individuals (neurodiversity) in both health and disease. Such progress has not yet, however,propelled changes in educational or medical practices that improve people’s lives. We review neuroimagingfindings in which initial brain measures (neuromarkers) are correlated with or predict future education,learning, and performance in children and adults; criminality; health-related behaviors; and responses topharmacological or behavioral treatments. Neuromarkers often provide better predictions (neuroprognosis),alone or in combination with other measures, than traditional behavioral measures. With further advances instudy designs and analyses, neuromarkers may offer opportunities to personalize educational and clinicalpractices that lead to better outcomes for people.

Noninvasive neuroimaging has provided remarkable new in-

sights into human brain structure and function in both health

and disease. For over a century, understanding the human brain

depended on naturally occurring brain injuries or unexpected

consequences of neurosurgeries. From clinical cases such as

Leborgne, Phineas Gage, H.M., and commissurotomy patients,

we gleaned insights, respectively, into the roles of the left pre-

frontal cortex in language (Broca, 1861), the ventral prefrontal

cortex in decision-making and social behavior (Harlow, 1868),

the medial temporal lobe in memory (Scoville and Milner,

1957), and functional asymmetries between the cerebral hemi-

spheres (Gazzaniga, 1970). Noninvasive neuroimaging has

permitted a second wave of discoveries about the brain that

has expanded the horizon of human neuroscience with examina-

tion of typical functions across many domains of the human

mind, from perception and cognition to emotion, social and

moral thought, and economic decision-making. Furthermore,

such imaging has offered the first compelling evidence that

neuropsychiatric and neurodevelopmental disorders reflect

fundamental differences in brain structure and function.

Uniquely, neuroimaging has revealed not only universal princi-

ples of functional brain organization but also neurodiversity:

how such brain functions vary across people in relation to age,

sex, personality, culture, and genetics. Here we review the prog-

ress in a novel application of neuroimaging: the use of such

measureable neurodiversity to predict future human behavior.

Such prediction may constitute a humanitarian and pragmatic

contribution of human cognitive neuroscience to society, but

this contribution will require rigorous science and also ethical

considerations.

Neuroscientists, psychologists, and physicians are contem-

plating how human neuroimaging may best inform basic and

clinical research. For basic research, there is discussion about

whether neuroimaging has informed cognitive theories beyond

the mapping of psychological functions to neural networks

(e.g., Mather et al., 2013). For clinical research, it is noteworthy

that the 2013 revision of the Diagnostic and Statistical Manual

of Mental Disorders (DSM-5), the defining document of diag-

nosis from the American Psychiatric Association, was little, if at

all, influenced by the over 15,000 MRI studies of psychiatric dis-

orders listed in PubMed (and this does not include studies using

other methods, such as electroencephalography (EEG), magne-

toencephalography (MEG), or positron emission tomography

(PET)). Remarkable advances in genetics have also had little

practical influence as yet on diagnosis or treatment of psychiatric

disorders. Because psychiatric disorders are known to be herita-

ble, and because these disorders must have a brain basis, it is

likely that progress in genetics and neuroimaging will illuminate

such disorders in the long run. Here we consider how neuroi-

maging may contribute to helping people in the nearer future.

This review focuses on structural and functional neuroimaging

and considers findings in which an initial brain measure (a neuro-

marker) is associated with a future behavioral outcome. Some

studies relate neuromarkers to individual differences in later

perceptual or cognitive performance among typical or healthy

people and have relevance for education and training. Other

studies relate neuromarkers to individual differences among

patients with a given diagnosis to future clinical status or

response to treatment (neuroprognosis) and have relevance for

neuropsychiatric disorders.

Such correlational or predictive studies differ from other kinds

of studies in two main ways. First, in the case of group studies

(e.g., comparison of patient and control groups), neuroimaging

differences aremost pronounced when there is greater homoge-

neity of a brain measure within each group so that groups are

statistically separable. Conversely, greater heterogeneity of a

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 11

Neuron

Review

brain measure within a group is more likely to yield neuromarkers

that correlate with variable outcomes. Second, for studies that

examine response to treatment, individual differencesmay delin-

eate not the neural systems most affected by the disorder but,

rather, heterogeneity among patients in the neural systems

that are most important and variable for how a treatment yields

benefits. For example, if a behavioral therapy helps a patient

with a disorder to learn how to regulate thoughts or emotions,

then the neuromarkers associated with treatment response

may be in neural networks that support such learning rather

than in networks related to the etiology or progression of the dis-

order.

Neuroimaging MeasuresNoninvasive neuroimaging measures provide indices of human

brain structure and function that vary in their strengths and limi-

tations. This review focuses on measures that maximize spatial

information, specifically MRI-derived measures. Brain structure

can be quantified by measuring volume, thickness, or density

(voxel-based morphometry [VBM]). Microstructural properties

of white matter pathways can be characterized by diffusion

tensor imaging (DTI). Brain functions can be quantified via fMRI

by activation studies that correlate experimental conditions or

behavioral performance with neural activity, as indexed by

changes in blood oxygenation level-dependent (BOLD) signals.

During a resting state with no task or stimuli, there are sponta-

neous fluctuations in functionally related brain regions that corre-

late with one another, and the patterns of these correlations may

reveal intrinsic functional relations of brain regions (Biswal et al.,

1995). Resting-state fMRI, EEG, and MEG can elucidate these

networks. Because it measures hemodynamic response, fMRI

is inherently poor in temporal resolution, whereas EEG and

MEG provide high temporal resolution (at the loss of spatial res-

olution).

For applications in education or medicine, there is a trade-off

between measures that are task-dependent (activation fMRI,

MEG, and EEG) versus measures that are task-independent

(structural MRI and DTI, fMRI, MEG, and EEG resting state).

On one hand, tasks can selectively invoke brain responses to

salient stimuli (e.g., to print in children with reading difficulty or

to sad facial expressions in depression). The advantage of this

approach is that tasks and stimuli can be tailored to specifically

assay salient mental operations. On the other hand, such tasks

demand participant performance that can result in behavioral

confounds, vary in design from study to study, and have not typi-

cally been developed to maximize reliability of measurement. In

contrast, structural and resting state measures can be acquired

in a consistent fashion, have promise for reliability (e.g., Shehzad

et al., 2009; Wonderlick et al., 2009; Vollmar et al., 2010), can

accommodate a broad range of participants (including infants),

and are independent of task performance in the scanner.

Analytic Approaches: FromCorrelation to IndividualizedPredictionAn ultimate goal of the use of neuromarkers for neuroprognosis

is to perform individualized predictions of educational or health

outcomes.Most studies to date have related variation in baseline

brain measures to variation in subsequent outcomes. Given that

12 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

such analyses hinge on knowledge of the outcomes, such ana-

lyses could be described more as postdiction than prediction

(Whelan and Garavan, 2014). However, if neuromarkers are to

become useful in practice, they must predict outcomes for

new individuals based on models developed previously with

other individuals. A cognitive neuroscience of prediction, there-

fore, needs to build on theory and methods that allow for effec-

tive creation, evaluation, and selection of prediction models

(Pereira et al., 2009).

The term prediction is used in three different ways in relevant

research. First, prediction can refer to a correlation between two

contemporaneous values, such as height predicting weight.

Second, prediction can refer to the correlation of one variable

in a group at an initial time point to another variable in the

same group at a future time point (an in-sample correlation).

Third, prediction can refer to a generalizable model that applies

to out-of-sample individuals. All studies reviewed here relate an

initial brainmeasure to a future behavioral outcome, and the term

correlation refers to in-sample findings, and the term prediction

refers to out-of-sample generalizations.

Such research can be conceptualized as comprising three

stages, beginningwith within-sample correlations to discover re-

lations of interest, progressing to predictive analyses in which

predictions for individuals are derived from data from other in-

sample individuals, and culminating in predictive analyses in

which a model from one sample is used to predict outcomes

in an independent sample (Figure 1). Each stage requires more

participants, so that prior stages may justify larger-scale studies.

The vastmajority of findings to date are correlational (61 of the 72

reviewed here), but some studies reported predictive analyses

(with only one study having fully independent samples) (Table 1).

The major limitation with correlational analyses reporting the

significance of the overall fit of linear or multiple regression

models to a data set is that findings are tied to the outcome for

a particular group. From a predictive modeling standpoint, the

error from this fit is typically termed the training error, whereas

the error on an unseen data set would be called the test or gener-

alization or prediction error. The training error is always an under-

estimate of the test error. The quality of amodel can be evaluated

by measuring its test error. Minimizing this error is the goal of

building predictionmodels. Oneway to decompose the test error

is to describe it as a sum of training error and optimism (Efron,

2004). Optimism is the difference between the test error, which

is always higher, and the training error.

The most common approach for reducing optimism is to use a

validation set in which some data are set aside to estimate the

test error. In many studies of brain imaging, this limits the amount

of data available for training because of small sample sizes. A

common approach is to use cross-validation, in which one di-

vides a data set into a number of folds. One fold is held aside

as a test set, and data from the remaining folds (training data)

are used to train the model. This model is then applied to the

test set, and the model error is calculated. This procedure is

repeated by considering each fold as a test set. The average er-

ror across the test folds is reported as the generalization error. If

the number of folds equals the number of data points, then only

one data point is held out for testing, and this is known as leave

one out cross-validation. In general, this approach is unbiased

Figure 1. Three Stages of Predictive ModelIdentification(1) Discovery phase. Explore and evaluate asso-ciations between baseline neuromarkers andbehavioral outcomes.(2) Cross-validation phase. A cross-validationroutine is used to separate data into training andtest sets. The model is built using training data andtested on out-of-sample test data. Upon suc-cessful evaluation of the performance of the modeland features, all data are used to build a predictionmodel.(3) Generalization phase. A prediction model builtvia cross-validation is applied to a new data set.The new data are then used to update the model.

Neuron

Review

but typically has a high variance in prediction error (Kohavi, 1995;

Rao et al., 2008).

Another practical approach is to randomly split data into

training and test sets (e.g., 10% of the data are in the test set).

The splitting is repeated several times. On each iteration, a

model is fit on the training data and tested on the test data.

This results in a distribution of prediction errors that can provide

a confidence interval for a given application. However, such pro-

cedures can still lead to increased optimism if models are cho-

sen or their parameters are tuned after peeking at the test results

(Koban et al., 2013). Selecting models and their parameters from

cross-validation on the training data can reduce such optimism.

The training data itself can be subjected to cross-validation and

subdivided into training and test sets to determine which model

is best suited for the training data. This procedure is called

‘‘nested cross-validation.’’ Because different models may be

selected for each cross-validation split, the most selected model

might be considered to be the ‘‘best’’ model. A variety of learning

models coupled with cross-validation has been used in brain im-

aging. These range from linear, multiple, and logistic regression

models to approaches such as support vector machines (SVM

[Vapnik, 1999]) or least absolute shrinkage and selection oper-

ator (LASSO [Tibshirani, 1996]), relevance vector machines

(Bishop and Tipping, 2000), and random forests (Breiman, 2001).

The difference between the amount of variability accounted for

by within-sample correlations and out-of-sample predictions is

rarely reported. Two within-sample correlational studies (Aharoni

et al., 2013; Demos et al., 2012) were reanalyzed by a different

investigator (R.A. Poldrack, personal communication), and the

outcome variance accounted for by the generalizable model

was far smaller than that for the within-sample correlation (see

Aharoni et al., 2014). Although inmost cases the predictivemodel

Neuron

results in a more conservative outcome

than thecorrelationalmodel, thedifference

varies across data sets. In all cases, how-

ever, predictive analyseswill be necessary

to translate correlational observations into

educational or clinical practice.

Future Learning and CognitivePerformance in AdultsVariation in initial neuromarkers has been

associated with subsequent learning or

cognitive performance, and, in most cases, these variations

occurred in the neural networks associated with the kind of

learning. Larger volumes of the striatum correlated with superior

video game skill learning (Erickson et al., 2010). This correlation

was specific to the dorsal striatum volume, did not extend to the

hippocampus or ventral striatum, and accounted for 23% of the

variance in learning. The importance of the striatum for such skill

learning is consistent with evidence that lesions of the striatum

impair skill learning (e.g., Heindel et al., 1989). Superior word

learning correlated with DTI measures of the left arcuate fascic-

ulus, a white matter pathway connecting major left hemisphere

language regions (Lopez-Barroso et al., 2013).

Brain differences in language-related neural systems have

also been related to variation in learning novel speech distinc-

tions not present in a person’s native language. Superior learning

was associated with anatomical differences, specifically greater

asymmetry (left to right) in parietal lobe volumes and higher white

matter density in left Heschl’s gyrus (Golestani et al., 2002,

2007). Larger anatomical structures in the language-dominant

left hemisphere may support the rapid temporal processing

needed to learn novel auditory distinctions that occur critically

in the first 30–50 ms of nonnative language sounds. Resting-

state functional connectivity has also been associated with vari-

ation in auditory language learning. Better learners of a nonnative

speech contrast exhibited greater functional connectivity (corre-

lation) than poor learners between inferior frontal and parietal re-

gions thought to be major components of the left hemisphere

language system (Ventura-Campos et al., 2013; other related

studies are reviewed in Zatorre, 2013).

Neuromarkers have also correlated with musical and visual

learning. For auditory learning of microtonal pitch discrimination

(with intervals smaller than typically used in musical scales),

85, January 7, 2015 ª2015 Elsevier Inc. 13

Table 1. Publications Cited in This Article that Used Cross-Validation Techniques to Measure Out-of-Sample Prediction Error

Publication Sample Size Application Cross-Validation Method Learning Model

Whelan et al., 2014 271 future adolescent alcohol misuse 10-fold logistic regression with elastic net

Hoeft et al., 2007 64 future reading skills LOO linear regression

Ullman et al., 2014 62 future working memory capacity LOO nu-SVR

Ball et al., 2014 48 response to CBT in GAD and PD out of bag random forest

Doehrmann et al., 2013 39 SAD stratified K-fold linear regression

Hoeft et al., 2011 25 future reading gains in dyslexia LOO linear SVC

Supekar et al., 2013 24 response to math tutoring 4-fold linear regression

Falk et al., 2010 20 persuasion-induced behavior

change

2-fold linear regression

Siegle et al., 2012 17 (cohort 1)

20 (cohort 2)

response to CBT for depression out of bag random forest

Bach et al., 2013 17 future reading skills LOO discriminant analysis

Costafreda et al., 2009 16 response to CBT for depression LOO PCA + linear SVM

LOO, Leave One Out; nu-SVR, nu-Support Vector Regression; CBT, cognitive behavioral therapy; GAD, generalized anxiety disorder; PD, panic dis-

order; SAD, social anxiety disorder; SVC, Support Vector Classification; PCA, Principal Component Analysis; SVM, Support Vector Machine.

Neuron

Review

individuals who, at baseline, exhibited higher slopes of fMRI acti-

vation in the bilateral auditory cortex to pitch-interval size ex-

hibited greater learning over a 2-week training period (Zatorre

et al., 2012). The higher slope of activation may reflect a finer-

grained cortical encoding of pitch information that potentiates

more rapid learning during training. People who were better at

learning to make fine visual discriminations had, at baseline,

stronger functional connectivity within portions of the visual cor-

tex and between the visual cortex and prefrontal association

areas (Baldassarre et al., 2012). These regions were also a sub-

set of the regions that were activated by the discrimination task

itself, suggesting that initial individual differences within the task-

evoked neural networks encouraged or discouraged effective

learning.

The studies mentioned above examined variation in learning

across individuals, but individuals also vary across time in their

performance and learning. Two fMRI studies exploited natural

fluctuations in resting-state BOLD signals in an attempt to distin-

guish brain states within an individual that were associated with

superior or inferior performance on vigilance and learning tasks.

In both studies, stimulus presentation was triggered via real-time

fMRI when BOLD signals in relevant brain regions were hypoth-

esized to be in optimal or suboptimal states. In the vigilance task,

an individual had better vigilance (faster reaction times) for the

appearance of an unpredictable visual target when, before the

appearance of the target, the BOLD signal was high in the sup-

plementary motor area (a region associated with motor planning)

and low in components of the default mode network (a network

that is more active during rest than most tasks and that has been

associated with internal self-reflection rather than external

perceptual attention) (Hinds et al., 2013). In the memory task,

an individual exhibited superior learning of scenes when BOLD

signals were lower before the appearance of a scene in the pos-

terior parahippocampal cortex, a region that is selectively

responsive to scenes (Yoo et al., 2012). Therefore, brain states

could be identified that predicted whether an individual was

ready to be vigilant or ready to learn.

14 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

Future Learning and Education in ChildrenReading and mathematics are the two foundations of education,

and, accordingly, the focus of school curricula from elementary

school through high school. The first major education experience

for children is learning to read in early school years, after which

they use those reading skills to learn all other subjects. Some

children (5%–17%) have developmental dyslexia, which is a

persistent difficulty in learning to read that is not explained by

sensory, cognitive, or motivational factors or lack of adequate

reading instruction (Shaywitz, 1998) and that is highly heritable

(Pennington and Gilger, 1996). The best understood psycholog-

ical cause of dyslexia is a weakness in phonological awareness,

the understanding that spoken words are composed of discrete

sounds (phonemes) that can be mapped onto letters or syllables

(graphemes) (Bradley and Bryant, 1978), although several other

putative causes have been identified (reviewed in Gabrieli, 2009).

Brain measures in infants have correlated with future success

or failure in language and reading years before explicit reading

instruction. Event-related potentials (ERPs), which are time-

locked changes in electrical activity measured with EEG scalp

electrodes, have revealed a risk for future language and reading

difficulties in newborns within hours or days of birth. These

studies typically involve infants from families with a history of lan-

guage or reading difficulty to increase the proportion of infants

who will progress to language and reading difficulty. ERP re-

sponses to speech sounds within 36 hr of birth discriminated

with over 81% accuracy those infants who would go on to

become dyslexic at age 8 (Molfese, 2000). Newborns, tested

within a week of birth, had ERPs in response to speech sounds

that correlated with language scores at 2.5, 3.5, and 5 years of

age (Guttorm et al., 2005).

Some studies have reported that neuroimaging measures

enhance or outperform traditional behavioral measures in fore-

casting children’s reading abilities in future months and years.

One study examined how children 8–12 years of age, identified

by their teachers as struggling readers, fared from the beginning

to the end of a school year in single-word decoding skills (the

Figure 2. Functional and Structural BrainMeasures Predicting EducationalOutcomes(A and B) fMRI predictor of reading gains indyslexia. Greater activation for a phonological taskin the right inferior frontal gyrus (Rt IFG) (A) pre-dicted greater gains in reading 2.5 years later indyslexic children (B). Each red circle is an individ-ual (based on Hoeft et al., 2011).(C and D) MRI predictor of math tutoring gains instudents. Greater gray matter volume of the right(R) hippocampus (C) predicted greater perfor-mance gains in students after 8 weeks of tutoring(D). Each blue circle is an individual (from Supekaret al., 2013).

Neuron

Review

ability to read aloud pseudowords on the basis of phoneme-

grapheme mapping rules) (Hoeft et al., 2007). At the beginning

of the school year, these children were evaluated with over a

dozen behavioral measures of reading and reading-related skills,

an fMRI task requiring rhyme judgments for pairs of printed

words, and a VBM analysis of anatomic gray and white matter

densities. The beginning-of-the-year behavioral measures ac-

counted for 65% of the variance in end-of-year scores, and the

brain measures accounted for 57% of that variance. The combi-

nation of behavioral and brain measures accounted for a signif-

icantly better 81% of the variance, demonstrating enhanced

forecasting of student reading skills across a school year.

Among children with dyslexia, there is considerable variation

in the degree to which individual children do or do not compen-

sate for their reading difficulty by closing the gap between their

actual and age-expected reading skills. A longitudinal study of

older children (mean age of 14 years) examined how behavioral

measures (17 tests of reading and reading-related skills), fMRI

activation for a word-rhyming task, and DTI indices of white

matter organization predicted which children, over the next

2.5 years, would compensate or persist in their reading difficulty

(Hoeft et al., 2011). None of the standard behavioral measures

Neuron

correlated with future reading gains, but

the brainmeasures did yield such correla-

tions (Figure 2). In combination, greater

activation in the right prefrontal cortex

(a region not typically engaged for

reading single words at this age) and

greater white matter organization of the

right superior longitudinal fasciculus pre-

dicted, with 72% accuracy, whether a

child would be in the compensated or

persistent group. Multivoxel pattern anal-

ysis of whole-brain fMRI activation, a

data-driven pattern classification anal-

ysis, yielded over 90% accuracy in classi-

fyingwhether a dyslexic child, at baseline,

would belong to the compensating or

persistent group 2.5 years later.

Longitudinal studies have also found

neuromarkers associated with future

reading skills in children who were not

selected on the basis of family history or

reading difficulty. In a 5-year longitudinal study, an auditory

ERP measure (hemispheric lateralization of late mismatch nega-

tivity) in prereading kindergarteners significantly improved the

forecasting of future reading performance in second, third, and

fifth grades in combination with prereading skills (Maurer et al.,

2009). Only the ERP measure (and not any behavioral measure)

correlated with future reading performance in fifth grade. A visual

ERP study with prereading kindergarteners also reported that

the combination of behavioral measures and both ERP and

fMRI responses to print explained up to 88% of the variance in

second-grade grade reading ability (Bach et al., 2013). These

studies suggest that neuromarkers in prereading kindergar-

teners may enhance the identification of children who will strug-

gle to read even before reading instruction begins in school. This

is important because current reading interventions are most

effective in young, beginning readers, and effective intervention

prior to reading failure may not only be more effective but also

spare children the sense of failure that often accompanies early

struggles in reading.

In older typical readers 9–15 years of age, fMRI activations in

response to a word-rhyming task were associated with nonword

reading skills up to 6 years in the future, with the specific

85, January 7, 2015 ª2015 Elsevier Inc. 15

Neuron

Review

locations of activations depending upon the child’s age (McNor-

gan et al., 2011). In younger children, greater activation in brain

regions associated with phonological recoding (e.g., the inferior

frontal gyrus) was associated with greater future reading skill,

whereas, for older children, greater activation in brain regions

associated with orthographic analysis of print (e.g., the fusiform

gyrus) was associated with lesser future gains. These findings

underscore how different developmental stages of learning to

read, perhaps transitioning from a younger gaining of skill in sin-

gle-word decoding (print-to-sound correspondence) to an older

mastering of fluent visual analysis of connected print, may invoke

relatively different components of the brain’s reading circuitry.

There is also considerable variation in how well children can

learn a second language. For native Chinese speakers around

age 10, greater activation in response to English words and non-

words in the left fusiform gyrus and left caudate correlated with

superior English word-reading levels a year later (Tan et al.,

2011). The putative visual word form area, which is highly

responsive to learned print, is located in the left fusiform gyrus

(Dehaene and Cohen, 2011). The leftward lateralization of neuro-

markers may have been related to properties of alphabetic lan-

guages such as English because there is evidence that variation

in microstructural properties of right hemisphere white matter

pathways correlated with initial learning of Mandarin Chinese in

young adults (Qi et al., 2014). The rightward lateralization of neu-

romarkers in native English speakers associated with future suc-

cessful initial language learning may reflect the tonal and visuo-

spatial properties, respectively, of spoken and written Mandarin

Chinese. Therefore, neuromarkers correlated with second lan-

guage learning may vary depending on the kinds of mental re-

sources needed to learn different kinds of languages.

Mathematical problem-solving skills are the foundation of later

performance in science and engineering. Academic skill in arith-

metic relies on multiple cognitive processes, including working

memory, the mental processes that support the maintenance

and manipulation of goal-relevant information over brief time pe-

riods (reviewed in Raghubar et al., 2010). In a longitudinal study,

children 6–16 years of age underwent behavioral testing (work-

ing memory, reasoning, and arithmetical abilities) and fMRI while

performing a visuospatial working memory task (Dumontheil and

Klingberg, 2012). Neuroimaging analyses focused on the intra-

parietal sulcus (IPS), a brain region associated with both visuo-

spatial working memory and numerical representation. The

working memory and reasoning measures were independent

predictors of arithmetical performance 2 years later. The magni-

tude of visuospatial working memory activation in the left IPS

also predicted future arithmetical performance. Combining the

neuroimaging and behavioral data more than doubled the accu-

racy of predicting future mathematical ability compared with the

use of only behavioral data.

The future growth of working memory ability in the same age

range has also been better predicted by a combination of neuro-

imaging and behavioral measures than by behavioral measures

alone (Ullman et al., 2014). Interestingly, although current work-

ing memory capacity correlated with activation in frontal and pa-

rietal regions, future capacity was best predicted by structural

and functional measures of the basal ganglia and thalamus. Spe-

cifically, greater activation in the caudate and thalamus and

16 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

greater fractional anisotropy (FA) of surrounding white matter

as measured by DTI predicted future growth in working memory

over the next 2 years.

There is increasing interest in improving the effectiveness of

learning through teaching that takes into account variation

among students. One study examined whether neuromarkers

could identify which children would benefit from a math tutoring

program for third graders (ages 8 and 9) that encouraged stu-

dents to shift from counting to fact retrieval as a basis for an arith-

metic problem-solving strategy (Supekar et al., 2013). Individual

differences in how much students benefitted from the tutoring

program did not correlate with baseline behavioral scores on

tests of intelligence (IQ), working memory, or mathematical abil-

ities. Conversely, at baseline, greater right hippocampal volume

and resting-state intrinsic functional connectivity between the

right hippocampus and prefrontal and striatal regions correlated

with future performance improvements (Figure 2).

Future CriminalityThe criminal justice system is rife with demands for predictions of

future behaviors as judgments are made about bail, sentencing,

and parole. The demonstrated inaccuracy of expert clinical judg-

ments (Monahan, 1981) has motivated the use of an actuarial

approach that estimates risk for future antisocial behavior based

on characteristics such as age, sex, criminal history, and drug

use (e.g., Yang et al., 2010). Building on evidence that impulsivity

(behavioral disinhibition) is a major risk factor for recidivism,

brain activations to an impulse control task (go/no-go task)

were examined in 96 male offenders who were then followed

longitudinally (Aharoni et al., 2013). The likelihood that an

offender would be rearrested over a 4-year period doubled

when, at baseline, the offender had low activation in the anterior

cingulate cortex, a region associated with cognitive control and

especially the resolution of cognitive conflict. Although the corre-

lation between baseline brain activation and future rearrest was

significant, there was no or weaker correlation for other predic-

tors (age, scores on a psychopathy checklist, lifetime substance

abuse, or behavioral error rate on the scanner task).

Future HealthStudies have examined whether neuromarkers are related to

future health-related behaviors such as alcohol abuse, drug

abuse, or unhealthy eating. Alcohol use by underage drinkers

is an important public health problem because such use in ado-

lescents is risky and also associated with life-long alcoholism.

Heavy or binge drinking is the primary source of preventable

morbidity and mortality for the more than 6 million American col-

lege students (Wechsler et al., 2002). Early onset of alcohol use

by age 12 is associated with numerous undesirable outcomes in

adolescence (Gruber et al., 1996), and initiation of drinking

before age 15, versus after age 20, quadruples the likelihood of

alcoholism (Grant and Dawson, 1997).

In a longitudinal study, 12- to 14-year-olds with little or no his-

tory of substance abuse performed a go/no-go task of response

inhibition while undergoing fMRI (Norman et al., 2011). About

4 years later, these adolescents were divided into two groups

who did or did not transition to heavy use of alcohol. Widespread

reductions in baseline activation, including in the prefrontal and

Neuron

Review

anterior cingulate cortices, were found in adolescents who later

transitioned to heavy alcohol use relative to those who did not.

Among adolescents 16–19 years of age with an ongoing history

of substance use disorders, those who exhibited less prefrontal

and greater parietal activation on the same task had higher levels

of substance use over the following 18 months (Mahmood et al.,

2013). Overall, the findings suggest that a relative weakness in

the recruitment of anterior brain regions that aremost associated

with cognitive control of behavior may be a predisposition for

early alcohol use or sustained substance abuse.

Adolescents who exhibited greater activation in response to

monetary reward in the basal ganglia were more likely to engage

in substance use (alcohol and drugs) a year later (Stice et al.,

2013). In contrast, those who were already using substances at

baseline exhibited lesser activation in the basal ganglia at base-

line. These findings indicate that reward systems of the basal

ganglia are also involved in substance abuse but that brain mea-

sures of future risk for substance use may be quite different than

brain measures reflecting the consequence of current use of

substances.

The largest study of future adolescent misuse of alcohol fol-

lowed nearly 700 adolescents and collected detailed histories,

personality measures, genetic information, structural and func-

tional MRI data, and cognitive performance measures (Whelan

et al., 2014). fMRI tasks examined inhibitory control, reward pro-

cessing, and facial expressions of emotion. In 271 of these ado-

lescents, a multidomain analysis was used to predict future

binge drinking. The most robust brain predictors of future binge

drinking came from the right precentral and bilateral superior

frontal gyri, with contributions from several structural (gray mat-

ter volume) and functional (inhibitory control and reward

outcome) features. In the predictive model, these brain mea-

sures were coupled with life events, personality measures, and

an anxiety sensitivity subscale of the substance use risk profile

scale. Any one feature in isolation had only a modest influence

on prediction, and many of the features predicting future misuse

were different from the features dissociating groups of binge

drinkers and nonbinge drinkers. Such a study highlights the mul-

tiple causal factors for substance abuse and the scale of data

needed to predict the future unfolding of such multifaceted pro-

cesses.

Healthy eating to avoid or reduce obesity is also amajor public

health concern. Neuroimaging studies have reported that fMRI

activations in response to food-related pictures forecast future

changes in body mass index (BMI) over the next 6–12 months.

Two studies examined the relation between baseline fMRI acti-

vations and weight gain over the following year in girls identified

as having body image concerns. Activations in response to

palatable food occurred in brain regions associated with reward

anticipation (e.g., regions of basal ganglia) or reward valuation

(e.g., the orbitofrontal cortex). In one case, dopamine-related ge-

netic variation interacted with blunted brain activation to corre-

late with elevated risk for future weight gain (Stice et al., 2010).

In another case, lateral orbitofrontal cortical activation during

initial orientation to appetizing food cues correlated with future

increases in BMI over a 1-year period (but behavioral patterns

of response did not correlate with BMI increases) (Yokum

et al., 2011).

Another study demonstrated the specificity of brain activa-

tions to food cues in relation to future weight gain (Demos

et al., 2012).Women arriving at college sawpictures of food, sex-

ual scenes, and control pictures during fMRI. At baseline and

again toward the end of the school year, the women’s weights

and self-reports of sexual behavior were measured. A greater

initial response in the reward-responsive nucleus accumbens

for food pictures correlated with greater BMI gains 6 months

later, whereas a greater initial response to sexual scenes corre-

lated with greater sexual desire and more sexual experiences

6 months later.

In a related study, college-aged women participated in an

fMRI study of brain responses to pictures of food and for

response inhibition on a go/no-go task, followed by experience

sampling via smartphone. Over the course of 1 week, they

were periodically asked to report their desire to eat food, at-

tempts to resist the temptation to eat, and whether or not and

how much they actually ate (Lopez et al., 2014). Greater nucleus

accumbens activation to food pictures correlated with greater

desires for food, more likelihood to give in to the temptation to

eat, and larger amounts eaten. Greater activation of the inferior

frontal gyrus during response inhibition was associated with

reduced surrender to temptation and less eating. Overall, these

studies suggest that an interplay between the response to food

cues that occurs in reward-sensitive striatal and orbitofrontal re-

gions and the response in cognitive control regions of the lateral

prefrontal cortex contributes to future healthy or unhealthy

eating patterns.

Another health-related behavior is the use of sunscreens for

protection against sunburn and some forms of skin cancer. In

one study, participants saw slides communicating the impor-

tance and proper application of sunscreen (Falk et al., 2010).

Participants also reported recent use of, intention to use, and

attitudes toward sunscreen. Greater activation in the medial

prefrontal cortex, a brain region associated with self-refer-

ence, correlated with changes in the use of sunscreen as

measured by an unexpected self-report 1 week later. Brain

measures accounted for about 25% of the change in use of

sunscreen above and beyond self-reported changes in atti-

tudes and intentions following presentation of the health infor-

mation during scanning. Activation in the medial prefrontal

cortex may broadly represent value because the magnitude

of activation in that brain region (and the striatum) in response

to individual consumer goods was associated with subse-

quent preferences for choosing those goods (Levy et al.,

2011).

Some studies have examined how brain function at one point

in time correlates with mental health outcomes at future time

points independent of treatment. For example, greater amyg-

dala activation to emotional facial expressions among patients

with depression correlated with reduced symptoms of depres-

sion 6 month later, controlling for initial depression severity and

medication status (Canli et al., 2005). In a memory paradigm

with negative pictures, greater baseline activation for success-

fully recalled pictures in the posterior cingulate cortex and

medial prefrontal cortex correlated with greater improvement

in depressive symptoms 18 months later (Foland-Ross et al.,

2014).

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 17

Neuron

Review

Future Response to TreatmentBiomarkers in general, and neuromarkers in particular, are not

used currently to predict treatment response for neuropsychi-

atric disorders despite considerable evidence that any specific

pharmacological or behavioral treatment is likely to be effective

for some patients but ineffective for a considerable number of

other patients. Measurement of treatment efficacy varies, but it

typically involves a patient report or clinician observation, often

via a structured interview or questionnaire. A highly effective

treatment results in remission, the absence of symptoms, or in

a substantial response, defined as an outcome in which the pa-

tient remains somewhat symptomatic but is much improved

(Frank et al., 1991).

Across many neuropsychiatric diagnoses, remission or sub-

stantial response occurs in about 50% of patients for a given

therapy. For depression, cognitive behavioral therapy (CBT) is

effective in 40%–60% of patients (Hollon et al., 2002), and selec-

tive serotonin reuptake inhibitors are effective in 40%–60% of

patients, although many patients who fail to respond to an initial

treatment will respond to another treatment or combination of

treatments (Souery et al., 2006). Similar 40%–60%success rates

for a given pharmacological or behavioral treatment have been

reported for generalized anxiety (Pollack et al., 2003), social anx-

iety disorder (Otto et al., 2000), and attention deficit hyperactivity

disorder (Wender, 1998; Biederman et al., 2010). This variability

in treatment response, which is not understood and not simply a

consequence of disease severity, suggests that there are clini-

cally important neurobiological differences among patients

sharing a diagnostic label so that a specific treatment will be

effective for some but not other patients.

To a remarkable degree, there is an absence of evidence

about which treatment is likely to be effective for a particular pa-

tient. Although patients often do benefit from a second or third

sort of attempted treatment, there is considerable human and

economic cost for delaying effective treatment for patients and

families who are often in crisis. The idea of personalized medi-

cine, that people vary in their response to treatments and that

more effective medicine can be practiced by knowing which

treatment is most likely to benefit a particular patient, has been

associated often with genetics. It seems plausible, however,

that quantitative brain measures may also reveal variation

among patients that provides an evidence-based rationale for

what treatment is most likely to help a particular patient among

currently available treatments.

Future Response to Pharmacological Treatment

Over 20 studies of depression have reported that pretreatment

neuroimaging measures can correlate with or predict clinical

improvement following pharmacological treatment (reviewed in

Pizzagalli, 2011; Fu et al., 2013). In one study, prior to treatment,

reduced subgenual anterior cingulate cortex (ACC) metabolism

was measured by PET in patients who subsequently responded

poorly to medicine relative to either healthy controls or patients

who responded well to medication and who exhibited greater-

than-normal metabolism (Mayberg et al., 1997). No clinical mea-

sure, such as depression severity, or behavioral measure, such

as cognitive performance, distinguished the patients who would

or would not respond to treatment. The subgenual ACC (Brod-

mann area 25) is especially salient for depression because it

18 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

has been shown to be functionally and structurally atypical in

depression (Drevets et al., 1997) and has been a target for

deep brain stimulation treatment of depression (Mayberg et al.,

2005).

A meta-analysis of 20 studies on depression supported the

conclusion that increased baseline activation of the ACC, ex-

tending into the orbitofrontal cortex, was associated with better

treatment response but that decreased activation of the insula

and striatumwas also associatedwith better treatment response

(Fu et al., 2013). In an fMRI activation study in which patients

viewed faces with sad facial expressions of varying intensity, a

machine learning approach (SVM and leave one out (LOO)

cross-validation) identified patients who would have remission

with 71% sensitivity/86% specificity (Costafreda et al., 2009).

There is also evidence that structural brain measures at baseline

were associated with treatment outcome. Across studies, a

worse response to treatment has been associated with

decreased gray matter volume in the left dorsolateral prefrontal

cortex and also in the right hippocampus (Fu et al., 2013). Finally,

repetitive transcranial magnetic stimulation (rTMS) is a less com-

mon treatment for depression, but resting-state functional con-

nectivity measures have been associated with clinical response

to such treatment (Salomons et al., 2014). Higher corticocortical

connectivity and lower corticothalamic, corticostriatal, and cor-

ticolimbic connectivity at baseline were associated with better

treatment response.

In an open-label study examining the efficacy of treating

generalized anxiety disorder with venlafaxine, patients viewed

faces with fearful or neutral expressions. Greater activations

for fearful relative to neutral faces in the rostral ACC and lesser

activations for fearful relative to neutral faces in the left amygdala

both correlated with greater clinical improvement (Whalen et al.,

2008). These correlations occurred despite no activation differ-

ences in the rostral ACC or amygdala either between patients

and controls or between pretreatment and posttreatment in pa-

tients who did improve clinically in response to treatment. Such a

finding underscores the idea that neuromarkers that are associ-

ated with treatment response need not reflect the same func-

tions as those related to etiology.

Future Response to Behavioral Treatment

Perhaps the best validated kind of behavioral treatment for

neuropsychiatric disorders is CBT, meta-analyses of which indi-

cate it to be effective for many disorders, including depression,

generalized anxiety disorder, panic disorder, and social anxiety

disorder (e.g., Butler et al., 2006; Hofmann et al., 2012). Multiple

studies have reported that CBT is similarly effective as pharma-

cological treatments for depression (DeRubeis et al., 2005),

generalized anxiety disorder (Mitte, 2005), pediatric anxiety

(Walkup et al., 2008), and social anxiety disorder (Heimberg

et al., 1998). For disorders that are treated primarily with medica-

tions, CBT has been shown to enhance clinical outcome relative

to other augmentations for obsessive-compulsive disorder

(OCD) (Simpson et al., 2013) and schizophrenia (Grant et al.,

2012).

Several neuroimaging studies have reported that pretreatment

neuroimaging measures correlate with or predict the magnitude

of clinical improvement following CBT in unipolar major depres-

sion, schizophrenia, and social anxiety disorder. The initial study

Figure 3. Functional Brain Measure Predicting a Clinical OutcomePrior to treatment, patients with social anxiety disorder who exhibited greaterposterior activation (left) for angry relative to neutral facial expressions had abetter clinical response to CBT than patients who exhibited lesser activation(right) (based on Doehrmann et al., 2013).

Neuron

Review

relating pretreatment brain function to clinical efficacy of CBT

occurred in 14 unmedicated patients with depression who

viewed emotionally negative words prior to treatment. Both

less sustained activation in the subgenual ACC and more sus-

tained activation in the amygdala were associated with a greater

improvement in response to CBT (Siegle et al., 2006). The finding

that less sustained activation in the subgenual ACC was associ-

ated with a better future response to CBT was replicated and

extended in a larger study of patients with depression (Siegle

et al., 2012). This study is noteworthy in its use of a model gener-

ated from one cohort being used to predict the outcomes of an

independent cohort.

For patients with schizophrenia being treated pharmacologi-

cally, about half respond beneficially to additional CBT treatment

(e.g., Wykes et al., 2008). In one set of overlapping studies, pa-

tients receiving CBT exhibited clinical improvements relative to

patients who did not receive CBT (Kumari et al., 2009, 2011; Pre-

mkumar et al., 2009). The magnitude of clinical benefit among

the patients receiving 6–8 months of CBT correlated with both

baseline functional and structural measures. Patients who ex-

hibited a stronger activation in the dorsolateral prefrontal cortex

(DLPFC) during performance of a working memory task and who

exhibited stronger DLPFC-cerebellar functional connectivity in

the most demanding condition of the task derived a greater

benefit fromCBT (Kumari et al., 2009). In another fMRI study, pa-

tients read aloud single words, heard either their own or another

person’s voice that was or was not distorted, and then judged

whether they had heard their own voice or that of another (Ku-

mari et al., 2010). Across several contrasts, greater activation

in the left inferior frontal gyrus and lesser inferior parietal and

medial prefrontal deactivation were associated with a greater

CBT benefit. A greater engagement of prefrontal regions in pa-

tients who benefitted more fromCBTmay be related to regulato-

ry processes that can support effective CBT. There has also

been some evidence for separable neuromarkers related to

CBT response for positive symptoms (excess or distorted

normal functions such as hallucinations or delusions) versus

negative symptoms (diminished normal functions such as apathy

or withdrawal) (Premkumar et al., 2009). Importantly, baseline

symptom severity did not correlate with CBT response so that

the neuromarkers provided measures associated with future

CBT benefits that were not clinically evident at baseline.

Current behavioral measures poorly predict treatment

outcome in social anxiety disorder, another disorder often

treated with CBT. Prior to CBT, patients viewed angry versus

neutral faces or negative versus neutral scenes during fMRI

(Doehrmann et al., 2013). Consistent with the social nature of

this anxiety disorder, activations in response to scenes were

not associated with treatment outcome, but activations to angry

relative to neutral faces were associated with CBT outcome.

Greater activations in higher-order visual cortices were predic-

tive of a superior treatment outcome (Figure 3). Initial greater

clinical severity accounted for about 12%of the variance in treat-

ment outcome, whereas the combination of baseline neuroimag-

ing and clinical severity accounted for about 40%of the outcome

variance. Similar findings were observed at a less conservative

statistical threshold in prefrontal cortices, and it is possible that

the relations between prefrontal and higher-order visuopercep-

tual cortices may support or constrain the self-regulatory pro-

cesses that are taught in CBT (i.e., that these results revealed

variation in the neural mechanisms that support CBT response

rather than those of social anxiety disorder).

A study of patients with generalized anxiety disorder or panic

disorder aimed to develop measures that might be sensitive to

single-subject responses to treatment (Ball et al., 2014). Patients

saw negative scenes and either maintained or reduced (via reap-

praisal) their emotional response to the scenes. A random forest

classification was used to identify brain regions in which activa-

tions best predicted treatment outcome. There were greater

activations for responders than nonresponders in the hippocam-

pus during the maintenance of negative images and in the ante-

rior insula and superior temporal, supramarginal, and superior

frontal gyri during reappraisal of negative images. The neuroi-

maging measures yielded superior accuracy, sensitivity, and

specificity in identifying individual patients as future responders

or nonresponders to treatment than clinical or demographic vari-

ables. This study provides an example of data-driven analyses

that are predictive even though the specific patterns of activation

are not readily interpretable in a cognitive neuroscience frame-

work.

For OCD, structural measures at baseline have been associ-

ated with variability in response to exposure therapy (Fullana

et al., 2014). A thinner cortex in the left rostral ACC at baseline

was associated with better responses to therapy. This same re-

gion was thinner overall in patients than in controls, so greater

differences from controls were associated with better outcomes.

The neuroanatomical locus is similar to that often observed in

studies of depression outcomes, which raises the possibility

that similar neural mechanisms may support behavioral thera-

pies across diagnoses.

The studies mentioned above examined the relations between

pretreatment neuromarkers and one kind of treatment, such as

CBT, a medication, or rTMS. The relevant choice that must be

made by a patient or physician, however, is not whether to pur-

sue one kind of treatment but, rather, to select among alternative

treatments. Therefore, an important and practical goal is to

examine whether there are differential predictors of effective-

ness for alternative treatments. One study employed PET imag-

ing prior to patients being randomly assigned to a medication

(escitalopram oxalate) or CBT to treat their depression (McGrath

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 19

Figure 4. Treatment-Specific Biomarker Candidates for Treatment of DepressionMean regional activity values for remitters and nonresponders segregated by treatment (either escitalopram given as escitalopram oxalate or CBT) are plotted forthe six regions showing a significant treatment 3 outcome analysis of variance interaction effect. Regional metabolic activity values are displayed as region/whole-brain metabolism converted to z scores (from McGrath et al., 2013b).

Neuron

Review

et al., 2013b). Six limbic and cortical regions showed a differen-

tial response to the two treatments, with right anterior insula hy-

pometabolism correlating with future remission to CBT (and poor

medication response) and right anterior insula hypermetabolism

correlating with future remission to medication (but a poor

response to CBT) (Figure 4). Subgenual ACC metabolism was

higher in patients who failed to respond to either treatment

than in patients who remitted from depression (McGrath et al.,

2013a).

Another study with a small number of pediatric patients with

generalized anxiety also found correlations between pretreat-

ment brain functions and treatment outcomes but did not find

differences between behavioral (CBT) and pharmacological

(fluoxetine) treatments. Greater baseline activation of the amyg-

dala to negative facial expressions was associated with better

symptom improvement regardless of treatment type (McClure

et al., 2007).

Future Relapse for Alcohol, Drug Addiction, Smoking,

and Diet

Alcoholism, drug addiction, smoking, and obesity aremajor pub-

lic health problems that share a similar treatment aim, namely

abstinence from a substance that is harmful to the brain and

body. Several studies have examined relations between neuro-

markers and whether individuals abstain successfully or relapse

into their health problems. Generally, these studies examined

patients who recently became abstinent at the initiation or termi-

nation of a treatment program and then followed these patients

over weeks or months to learn which patients continued to

abstain versus those who relapsed. Improved identification of

risk for relapse could support individualized treatment ap-

20 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

proaches that vary for those at minimal or maximal risk for

relapse.

At least 60%of patients who seek treatment for an alcohol use

disorder relapse within 6 months following treatment (Maisto

and Connors, 2006; Udo et al., 2009), and there have been

several studies in which baseline brain measures are associated

with future abstinence versus relapse. In two studies of recently

abstinent patients, greater activation of the putamen was asso-

ciated with a greater likelihood of relapse 3 weeks or 3 months

later (Braus et al., 2001; Grusser et al., 2004). Other measures,

such as self-reported intensity of craving, history of intake, or

duration of abstinence before scanning, were not associated

with likelihood of relapse. Both anatomic (Rando et al., 2011)

and regional cerebral blood flow (Noel et al., 2002) studies re-

ported that baseline measures of the medial prefrontal cortex

were associated with likelihood of relapse. Similarly, patients

who relapsed exhibited reduced volumes of the medial and/or

lateral prefrontal cortex (Durazzo et al., 2011; Cardenas et al.,

2011) and reduced white matter FA in frontal regions (Sorg

et al., 2012) relative to patients who sustained abstinence.

Broadly, a greater reward response to alcohol-related stimuli

and lesser strength in cognitive control regions were related to

relapse.

Relapse after treatment occurs at an estimated 50% rate

within a year among individuals with stimulant dependence

who seek treatment (Miller, 1996). Several neuroimaging studies

have reported that neuromarkers can contribute to the identifi-

cation of future abstinence or relapse. One group of patients

with methamphetamine dependence underwent fMRI 3 or

4 weeks after cessation of drug use and near completion of a

Neuron

Review

28-day inpatient program and were followed for about a year, at

which point about half of the patients had relapsed (Paulus et al.,

2005). During fMRI, participants attempted to either predict

where a stimulus would appear or to simply note that a stimulus

had appeared. None of the multiple sociodemographic, baseline

symptom, or use characteristics predicted relapse, but patients

who would later relapse exhibited a greater activation in multiple

brain regions than those who would not relapse, including the

prefrontal, parietal, and insular cortices. A pattern of activation

across right insular, posterior cingulate, and temporal regions

correctly identified 20 of 22 patients who did not relapse and

17 of 18 patients who did relapse. Other studies have reported

correlations between baseline fMRI activations and future

relapse for cocaine use after 1 week (Prisciandaro et al.,

2013), after a 10-week outpatient program (and a better predic-

tor than subjective reports of craving) (Kosten et al., 2006), and

after an 8-week outpatient program (Brewer et al., 2008; Jia

et al., 2011). Specific locations of activations that correlated

with future relapse varied across these studies, perhaps reflect-

ing differences in experimental paradigms, analyses, or partici-

pants.

Tobacco smoking is the leading preventable cause of death in

the developed world, with one billion tobacco-related deaths

projected for the 21st century (World Health Organization,

2008). Identifying smokers at high risk for relapse could influence

the design of cessation programs to fit with individual risk pro-

files. In one study, adult nicotine-dependent women underwent

fMRI while viewing smoking-related and unrelated pictures

before quitting smoking (Janes et al., 2010). The women then

made an attempt to quit smoking, and prequit measures were

related to subsequent success or failure in smoking cessation.

A greater baseline activation to smoking-related pictures in the

insula correlated with the likelihood of future relapse. The identi-

fication of insula reactivity as a correlate of future relapse is of in-

terest because lesions to the insula in smokers were associated

with reduced smoking that was immediate and without relapse

(Naqvi et al., 2007). Smokers who did not quit successfully also

exhibited reduced functional connectivity between an insula-

containing network and dorsolateral prefrontal cortex and dorsal

ACC, suggesting a weakness in interactions between brain re-

gions associated with smoking desires with regions associated

with cognitive control. A combination of brain functional data

and a behavioral task resulted in 79% accuracy in identifying

smokers who would or would not abstain from smoking. Future

success in quitting smoking has also been associated with

graymatter volumes in cortical and subcortical regions (Froeliger

et al., 2010).

Brain measures may also help identify what sort of information

presented to people aiming to quit smoking are likely to be effec-

tive. Ads aimed at encouraging people about to try to quit smok-

ing were presented during fMRI, and relapse was followed for a

month (Falk et al., 2011). A greater activation in the medial pre-

frontal cortex at baseline was associated with successful quit-

ting. The addition of the brain measures to other measures

(self-reported intentions, self-efficacy, and ability to relate to

the ads) more than doubled the accuracy of a model accounting

for changes in smoking behavior. In another study with a large

number of smokers, increased activation in brain regions associ-

ated with self-reference, especially the medial prefrontal cortex,

in response to individually tailored smoking cessation messages

was associatedwith a better probability of quitting 4months later

(Chua et al., 2011).

Healthy eating is a goal for individuals with obesity, and there is

evidence that brain measures at baseline are associated with

short- and long-term outcomes in a weight loss program (Mur-

daugh et al., 2012). Obese individuals viewed high-calorie food

versus control pictures before and after a 12-week weight loss

program, with a 9-month follow-up. Greater baseline activation

in the nucleus accumbens, insula, and ACC in response to

high-calorie food pictures correlated with lesser weight loss after

12 weeks. Further, less successful weight maintenance at

9 months correlated with greater posttreatment activation in

the insula, ventral tegmental area, and other regions. The rele-

vant regions are associated with interoception (insula), reward

(nucleus accumbens and ventral tegmental area), and cognitive

control (anterior cingulate), which are all processes related to di-

etary decisions.

Future Response to Placebos

Positive medical responses to placebo treatments are often

powerful and can rival the effectiveness of active treatments,

such as medicines for depression (e.g., Walsh et al., 2002).

Consequently, exploitation of placebo mechanisms may offer a

safe therapeutic approach for some patients, but there is evi-

dence for considerable variation in response to placebos (Walsh

et al., 2002). Most studies of the brain basis of individual differ-

ences in placebo responses have focused on pain, in part

because cortical and subcortical brain regions involved in pain

have been relatively well characterized. Placebo analgesia (the

positive influence of a placebo on experienced pain) was related

to a pattern of increased activation in several cortical control re-

gions and decreased activation in somatosensory activation dur-

ing the anticipation of pain rather than activation during reported

analgesia to pain itself (Wager et al., 2011). Patients with a better

future response to placebo treatments exhibited lesser resting-

state functional connectivity between the medial prefrontal cor-

tex and insula during a pain-rating task (Hashmi et al., 2012).

Furthermore, greater network efficiency during the resting state

was associated with a better response to future psychologically

mediated analgesia related to treatment for chronic knee pain

(Hashmi et al., 2014). A range of other findings also indicates

that functional and structural brain measures may help identify

individual patients most likely to benefit from placebo treatments

(reviewed by Koban et al., 2013).

Predicting Individual Futureswith Neuromarkers: Hopesand ChallengesAs reviewed above, neuromarkers obtained from noninvasive

brain imaging have shown great promise for identifying children

and adults more likely to learn well or poorly in particular

domains; more likely to progress to unhealthy (or even criminal)

behaviors; and more likely to respond to particular pharmaco-

logical, behavioral, or placebo treatments for many neurodeve-

lopmental and neuropsychiatric disorders. Although the

amount of scientific evidence is modest in many areas (with

reading and depression having perhaps the greatest concen-

trations of studies to date), there are also numerous studies

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 21

Neuron

Review

reporting that predictive neuromarkers either outperform or

significantly enhance traditional measures of individual vari-

ability, such as self-reports, clinical rating scales, or scores

on educational or neuropsychological tests. It is these kinds

of studies that express both a practical and humanitarian pos-

sibility of improving lives through recognizing individual differ-

ences in brain function and structure that greatly influence

the diversity of educational and clinical outcomes and using

that recognition to individually optimize educational and clinical

practices.

Because of these hopes, the challenges of translating cogni-

tive neuroscience measures into better futures for people need

to be carefully identified and thoughtfully overcome. First,

many of the reviewed studies were performed with relatively

small samples, and, in particular, many of the older studies

used statistical approaches that were overly liberal by current

standards. Although such pioneering studies must often begin

with modest resources because it is their outcomes that justify

larger studies, the translation of such science to practical appli-

cation now requires larger studies that can support more

rigorous statistics. This is particularly true for studies of neurodi-

versity that focus on individual differences because there must

be adequate sampling not only of a population as a whole but

also of the diversity of individuals within that population. Second,

studies must mature from correlations between baseline mea-

sures and clinical or educational outcomes to predictive models

that apply the outcomes from one group (training set) to another

group (test set) and, finally, to an individual. This is essential

because use of such measures must operate with new individ-

uals for whom a clinical or educational intervention is being

planned. Third, few studies have integrated findings across mul-

tiple imaging modalities, even when the multiple brain measures

could be made during a single MRI session. Combining multiple

kinds of neuromarkers may enhance their predictive accuracy.

Fourth, it will be important for future studies to involve plausible,

alternative interventions (e.g., McGrath et al., 2013b) because

the question is less often whether a person should receive help

but, rather, which kind of help is most likely to rapidly improve

the person’s education, skills, or health.

Neuromarkers will be useful to the extent that they outper-

form, alone or in combination with traditional measures, mea-

sures that are otherwise available. Indeed, multiple studies

have reported such value from neuromarkers, but other studies

have not examined whether the neuromarkers significantly

improve predictions above and beyond readily available mea-

sures. All forms of brain and behavioral assessment improve

over time, and perhaps a new behavioral assessment will

outperform neuromarkers in the near or distant future. At the

same time, behavioral assessments in many educational and

clinical areas have been developed and maintained over many

years, so it is unknown when breakthroughs might occur.

Perhaps neuroimaging measures will also be useful tools in

developing a new generation of brain-validated behavioral as-

sessments that can be readily used in schools, hospitals, and

medical offices. At the conceptual limit, there ought to be a

strong relation between measures of mind and brain so that a

new generation of behavioral measures could capitalize on the

novel insights of neuroimaging.

22 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

If neuroimaging remains necessary for optimal prediction,

there could be concerns about the cost and availability of

MRIs or other measures. In this regard, the cost of MRI imag-

ing in particular has raised concerns about its potential wider

use. One solution for availability could be to use more trans-

portable technologies, such as wireless EEG devices, with

assessment paradigms developed through coupled MRI and

EEG studies. Any economic analysis, however, ought to

include the costs of current practices in which patients are

often inadvertently directed to treatments that turn out to be

ineffective for that patient (often around half of patients for a

given treatment in many cases) or in which children must

demonstrate academic failure before receiving educational

intervention. The cost of a neuropsychological assessment

and report for an individual child or adult, for example, often

exceeds that of an MRI.

If neuromarkers are proven to enhance prediction, there will be

ethical and societal issues to consider. Because of their biolog-

ical nature, brain measures can be overly valued and potentially

divert public and scientific interest in behavioral and social fac-

tors (Kagan, 2013). If neuromarkers become more useful, they

will provoke questions about how to most ethically use predic-

tive information to help people rather than simply select people

most likely to succeed. This important concern, however, must

be weighed against the questionable validity of many current

practices, such as the finding that parole decisions made by

experienced judges appear to be greatly influenced by the

time of day and proximity to a meal at which a case is reviewed

(Danziger et al., 2011) or that medical schools continue to

conduct interviews for admissions despite evidence that deci-

sions based on such interviews have no correlation with objec-

tive measures of medical school performance (DeVaul et al.,

1987; Milstein et al., 1981). For in-patient treatments for sub-

stance abuse, there is little scientific justification for the proto-

typical 28-day treatment period. Even an imperfect predictive

measure of relapse may lead to more rational treatment dura-

tions that are related to individual variation. Neuromarkers and

neuroprognosis may offer practical and valuable contributions

because so many current educational and medical decisions

occur in the absence of scientific evidence that can guide those

decisions.

This review considered mostly studies with relatively short-

term longitudinal educational and clinical outcomes, but future

research may also attempt to predict longer-term outcomes.

Educational and medical practices often respond to crisis,

such as failure in learning to read or coping with depression.

Longer-term research may examine whether neuromarkers

can help identify children at early risk with the hopes of divert-

ing those children away from a trajectory toward failure and

crisis, such as ongoing studies attempting to identify whether

infants at familial risk for autism will or will not progress to

autism over the next few years (Bosl et al., 2011; Wolff

et al., 2012). Such early predictions may require novel forms

of intervention (e.g., language learning remediation in 2- or

3-year-olds that minimizes their difficulty in learning to read

as 5- and 6-year-olds) with the hope that such children never

experience the crises as children or adults that now initiate

intervention.

Neuron

Review

ACKNOWLEDGMENTS

We thank Cynthia Gibbs, Sara Beach, Elizabeth Norton, Allyson Mackey, JonWalters, and Maheen Shermohammed for help with the manuscript. Writing ofthis paper was supported by the Poitras Center for Affective DisordersResearch at the McGovern Institute for Brain Research and by NIH GrantR01HD067312.

REFERENCES

Aharoni, E., Vincent, G.M., Harenski, C.L., Calhoun, V.D., Sinnott-Armstrong,W., Gazzaniga, M.S., and Kiehl, K.A. (2013). Neuroprediction of future rearrest.Proc. Natl. Acad. Sci. USA 110, 6223–6228.

Aharoni, E., Mallett, J., Vincent, G.M., Harenski, C.L., Calhoun, V.D., Sinnott-Armstrong, W., Gazzaniga, M.S., and Kiehl, K.A. (2014). Predictive accuracyin the neuroprediction of rearrest. Soc. Neurosci. 9, 332–336.

Bach, S., Richardson, U., Brandeis, D., Martin, E., and Brem, S. (2013). Print-specific multimodal brain activation in kindergarten improves prediction ofreading skills in second grade. Neuroimage 82, 605–615.

Baldassarre, A., Lewis, C.M., Committeri, G., Snyder, A.Z., Romani, G.L., andCorbetta, M. (2012). Individual variability in functional connectivity predictsperformance of a perceptual task. Proc. Natl. Acad. Sci. USA 109, 3516–3521.

Ball, T.M., Stein, M.B., Ramsawh, H.J., Campbell-Sills, L., and Paulus, M.P.(2014). Single-subject anxiety treatment outcome prediction using functionalneuroimaging. Neuropsychopharmacology 39, 1254–1261.

Biederman, J., Mick, E., Surman, C., Doyle, R., Hammerness, P., Kotarski, M.,and Spencer, T. (2010). A randomized, 3-phase, 34-week, double-blind, long-term efficacy study of osmotic-release oral system-methylphenidate in adultswith attention-deficit/hyperactivity disorder. J. Clin. Psychopharmacol. 30,549–553.

Bishop, C.M., and Tipping, M.E. (2000). Variational relevance vectormachines.Proceedings of the Sixteenth Conference on Uncertainty in Artificial Intelli-gence (Morgan Kaufmann Publishers), pp. 46–53.

Biswal, B., Yetkin, F.Z., Haughton, V.M., and Hyde, J.S. (1995). Functionalconnectivity in the motor cortex of resting human brain using echo-planarMRI. Magn. Reson. Med. 34, 537–541.

Bosl, W., Tierney, A., Tager-Flusberg, H., and Nelson, C. (2011). EEGcomplexity as a biomarker for autism spectrum disorder risk. BMC Med. 9,18–34.

Bradley, L., and Bryant, P.E. (1978). Difficulties in auditory organisation as apossible cause of reading backwardness. Nature 271, 746–747.

Braus, D.F., Wrase, J., Grusser, S., Hermann, D., Ruf, M., Flor, H., Mann, K.,and Heinz, A. (2001). Alcohol-associated stimuli activate the ventral striatumin abstinent alcoholics. J. Neural Transm. 108, 887–894.

Breiman, L. (2001). Random Forests. Mach. Learn. 45, 5–32.

Brewer, J.A., Worhunsky, P.D., Carroll, K.M., Rounsaville, B.J., and Potenza,M.N. (2008). Pretreatment brain activation during stroop task is associatedwith outcomes in cocaine-dependent patients. Biol. Psychiatry 64, 998–1004.

Broca, P. (1861). Remarks on the seat of the faculty of articulate language, fol-lowed by an observation of aphemia. In Some Papers on the Cerebral Cortex(Charles C. Thomas), pp. 49–72.

Butler, A.C., Chapman, J.E., Forman, E.M., and Beck, A.T. (2006). The empir-ical status of cognitive-behavioral therapy: a review of meta-analyses. Clin.Psychol. Rev. 26, 17–31.

Canli, T., Cooney, R.E., Goldin, P., Shah, M., Sivers, H., Thomason, M.E.,Whit-field-Gabrieli, S., Gabrieli, J.D., and Gotlib, I.H. (2005). Amygdala reactivity toemotional faces predicts improvement in major depression. Neuroreport 16,1267–1270.

Cardenas, V.A., Durazzo, T.C., Gazdzinski, S., Mon, A., Studholme, C., andMeyerhoff, D.J. (2011). Brain morphology at entry into treatment for alcoholdependence is related to relapse propensity. Biol. Psychiatry 70, 561–567.

Chua, H.F., Ho, S.S., Jasinska, A.J., Polk, T.A., Welsh, R.C., Liberzon, I., andStrecher, V.J. (2011). Self-related neural response to tailored smoking-cessa-tion messages predicts quitting. Nat. Neurosci. 14, 426–427.

Costafreda, S.G., Khanna, A., Mourao-Miranda, J., and Fu, C.H. (2009). Neuralcorrelates of sad faces predict clinical remission to cognitive behavioural ther-apy in depression. Neuroreport 20, 637–641.

Danziger, S., Levav, J., and Avnaim-Pesso, L. (2011). Extraneous factors injudicial decisions. Proc. Natl. Acad. Sci. USA 108, 6889–6892.

Dehaene, S., andCohen, L. (2011). The unique role of the visual word form areain reading. Trends Cogn. Sci. 15, 254–262.

Demos, K.E., Heatherton, T.F., and Kelley, W.M. (2012). Individual differencesin nucleus accumbens activity to food and sexual images predict weight gainand sexual behavior. J. Neurosci. 32, 5549–5552.

DeRubeis, R.J., Hollon, S.D., Amsterdam, J.D., Shelton, R.C., Young, P.R.,Salomon, R.M., O’Reardon, J.P., Lovett, M.L., Gladis, M.M., Brown, L.L.,and Gallop, R. (2005). Cognitive therapy vs medications in the treatment ofmoderate to severe depression. Arch. Gen. Psychiatry 62, 409–416.

DeVaul, R.A., Jervey, F., Chappell, J.A., Caver, P., Short, B., and O’Keefe, S.(1987). Medical school performance of initially rejected students. JAMA 257,47–51.

Doehrmann, O., Ghosh, S.S., Polli, F.E., Reynolds, G.O., Horn, F., Keshavan,A., Triantafyllou, C., Saygin, Z.M., Whitfield-Gabrieli, S., Hofmann, S.G., et al.(2013). Predicting treatment response in social anxiety disorder from func-tional magnetic resonance imaging. JAMA Psychiatry 70, 87–97.

Drevets, W.C., Price, J.L., Simpson, J.R., Jr., Todd, R.D., Reich, T., Vannier,M., and Raichle, M.E. (1997). Subgenual prefrontal cortex abnormalities inmood disorders. Nature 386, 824–827.

Dumontheil, I., and Klingberg, T. (2012). Brain activity during a visuospatialworking memory task predicts arithmetical performance 2 years later. Cereb.Cortex 22, 1078–1085.

Durazzo, T.C., Tosun, D., Buckley, S., Gazdzinski, S., Mon, A., Fryer, S.L., andMeyerhoff, D.J. (2011). Cortical thickness, surface area, and volume of thebrain reward system in alcohol dependence: relationships to relapse andextended abstinence. Alcohol. Clin. Exp. Res. 35, 1187–1200.

Efron, B. (2004). The estimation of prediction error. J. Am. Stat. Assoc. 99,619–632.

Erickson, K.I., Boot, W.R., Basak, C., Neider, M.B., Prakash, R.S., Voss, M.W.,Graybiel, A.M., Simons, D.J., Fabiani, M., Gratton, G., and Kramer, A.F. (2010).Striatal volume predicts level of video game skill acquisition. Cereb. Cortex 20,2522–2530.

Falk, E.B., Berkman, E.T., Mann, T., Harrison, B., and Lieberman, M.D. (2010).Predicting persuasion-induced behavior change from the brain. J. Neurosci.30, 8421–8424.

Falk, E.B., Berkman, E.T., Whalen, D., and Lieberman, M.D. (2011). Neural ac-tivity during health messaging predicts reductions in smoking above andbeyond self-report. Health Psychol. 30, 177–185.

Foland-Ross, L.C., Hamilton, P., Sacchet, M.D., Furman, D.J., Sherdell, L.,and Gotlib, I.H. (2014). Activation of the medial prefrontal and posterior cingu-late cortex during encoding of negative material predicts symptom worseningin major depression. Neuroreport 25, 324–329.

Frank, E., Prien, R.F., Jarrett, R.B., Keller, M.B., Kupfer, D.J., Lavori, P.W.,Rush, A.J., and Weissman, M.M. (1991). Conceptualization and rationale forconsensus definitions of terms in major depressive disorder. Remission, re-covery, relapse, and recurrence. Arch. Gen. Psychiatry 48, 851–855.

Froeliger, B., Kozink, R.V., Rose, J.E., Behm, F.M., Salley, A.N., and McCler-non, F.J. (2010). Hippocampal and striatal gray matter volume are associatedwith a smoking cessation treatment outcome: results of an exploratory voxel-based morphometric analysis. Psychopharmacology (Berl.) 210, 577–583.

Fu, C.H., Steiner, H., and Costafreda, S.G. (2013). Predictive neural bio-markers of clinical response in depression: a meta-analysis of functional andstructural neuroimaging studies of pharmacological and psychological thera-pies. Neurobiol. Dis. 52, 75–83.

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 23

Neuron

Review

Fullana, M.A., Cardoner, N., Alonso, P., Subira, M., Lopez-Sola, C., Pujol, J.,Segalas, C., Real, E., Bossa, M., Zacur, E., et al. (2014). Brain regions relatedto fear extinction in obsessive-compulsive disorder and its relation to exposuretherapy outcome: a morphometric study. Psychol. Med. 44, 845–856.

Gabrieli, J.D. (2009). Dyslexia: a new synergy between education and cognitiveneuroscience. Science 325, 280–283.

Gazzaniga, M.S. (1970). The Bisected Brain. (Appleton-Century-Crofts).

Golestani, N., Paus, T., and Zatorre, R.J. (2002). Anatomical correlates oflearning novel speech sounds. Neuron 35, 997–1010.

Golestani, N., Molko, N., Dehaene, S., LeBihan, D., and Pallier, C. (2007). Brainstructure predicts the learning of foreign speech sounds. Cereb. Cortex 17,575–582.

Grant, B.F., and Dawson, D.A. (1997). Age at onset of alcohol use and its as-sociation with DSM-IV alcohol abuse and dependence: results from the Na-tional Longitudinal Alcohol Epidemiologic Survey. J. Subst. Abuse 9, 103–110.

Grant, P.M., Huh, G.A., Perivoliotis, D., Stolar, N.M., and Beck, A.T. (2012).Randomized trial to evaluate the efficacy of cognitive therapy for low-func-tioning patients with schizophrenia. Arch. Gen. Psychiatry 69, 121–127.

Gruber, E., DiClemente, R.J., Anderson, M.M., and Lodico, M. (1996). Earlydrinking onset and its association with alcohol use and problem behavior inlate adolescence. Prev. Med. 25, 293–300.

Grusser, S.M., Wrase, J., Klein, S., Hermann, D., Smolka, M.N., Ruf, M.,Weber-Fahr, W., Flor, H., Mann, K., Braus, D.F., and Heinz, A. (2004). Cue-induced activation of the striatum and medial prefrontal cortex is associatedwith subsequent relapse in abstinent alcoholics. Psychopharmacology(Berl.) 175, 296–302.

Guttorm, T.K., Leppanen, P.H., Poikkeus, A.M., Eklund, K.M., Lyytinen, P., andLyytinen, H. (2005). Brain event-related potentials (ERPs) measured at birthpredict later language development in children with and without familial riskfor dyslexia. Cortex 41, 291–303.

Harlow, J.M. (1868) Recovery after severe injury to the head. Publication of theMassachusetts Medical Society 2, 327–346.

Hashmi, J.A., Baria, A.T., Baliki, M.N., Huang, L., Schnitzer, T.J., and Apkarian,A.V. (2012). Brain networks predicting placebo analgesia in a clinical trial forchronic back pain. Pain 153, 2393–2402.

Hashmi, J.A., Kong, J., Spaeth, R., Khan, S., Kaptchuk, T.J., and Gollub, R.L.(2014). Functional network architecture predicts psychologically mediatedanalgesia related to treatment in chronic knee pain patients. J. Neurosci. 34,3924–3936.

Heimberg, R.G., Liebowitz, M.R., Hope, D.A., Schneier, F.R., Holt, C.S., Wel-kowitz, L.A., Juster, H.R., Campeas, R., Bruch, M.A., Cloitre, M., et al. (1998).Cognitive behavioral group therapy vs phenelzine therapy for social phobia:12-week outcome. Arch. Gen. Psychiatry 55, 1133–1141.

Heindel, W.C., Salmon, D.P., Shults, C.W., Walicke, P.A., and Butters, N.(1989). Neuropsychological evidence for multiple implicit memory systems: acomparison of Alzheimer’s, Huntington’s, and Parkinson’s disease patients.J. Neurosci. 9, 582–587.

Hinds, O., Thompson, T.W., Ghosh, S., Yoo, J.J., Whitfield-Gabrieli, S., Trian-tafyllou, C., and Gabrieli, J.D. (2013). Roles of default-mode network andsupplementary motor area in human vigilance performance: evidence fromreal-time fMRI. J. Neurophysiol. 109, 1250–1258.

Hoeft, F., Meyler, A., Hernandez, A., Juel, C., Taylor-Hill, H., Martindale, J.L.,McMillon, G., Kolchugina, G., Black, J.M., Faizi, A., et al. (2007). Functionaland morphometric brain dissociation between dyslexia and reading ability.Proc. Natl. Acad. Sci. USA 104, 4234–4239.

Hoeft, F., McCandliss, B.D., Black, J.M., Gantman, A., Zakerani, N., Hulme, C.,Lyytinen, H., Whitfield-Gabrieli, S., Glover, G.H., Reiss, A.L., and Gabrieli, J.D.(2011). Neural systems predicting long-term outcome in dyslexia. Proc. Natl.Acad. Sci. USA 108, 361–366.

Hofmann, S.G., Asnaani, A., Vonk, I.J., Sawyer, A.T., and Fang, A. (2012). Theefficacy of cognitive behavioral therapy: A review of meta-analyses. Cognit.Ther. Res. 36, 427–440.

24 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

Hollon, S.D., Thase, M.E., and Markowitz, J.C. (2002). Treatment and preven-tion of depression. Psychol. Sci. Public Interest 3, 39–77.

Janes, A.C., Pizzagalli, D.A., Richardt, S., deB Frederick, B., Chuzi, S., Pachas,G., Culhane, M.A., Holmes, A.J., Fava, M., Evins, A.E., and Kaufman, M.J.(2010). Brain reactivity to smoking cues prior to smoking cessation predictsability to maintain tobacco abstinence. Biol. Psychiatry 67, 722–729.

Jia, Z., Worhunsky, P.D., Carroll, K.M., Rounsaville, B.J., Stevens, M.C., Pearl-son, G.D., and Potenza, M.N. (2011). An initial study of neural responses tomonetary incentives as related to treatment outcome in cocaine dependence.Biol. Psychiatry 70, 553–560.

Kagan, J. (2013). Equal time for psychological and biological contributions tohuman variation. Rev. Gen. Psychol. 17, 351–357.

Koban, L., Ruzic, L., and Wager, T.D. (2013). Brain predictors of individual dif-ferences in placebo responding. In Placebo and Pain: From Bench to Bedside,L. Colloca, M.A. Flaten, and K. Meissner, eds. (Elsevier/Academic Press).

Kohavi, R. (1995). A study of cross-validation and bootstrap for accuracy esti-mation and model selection. In Proceedings of the 14th International JointConference on Artificial Intelligence (Morgan Kaufmann Publishers), pp.1137–1143.

Kosten, T.R., Scanley, B.E., Tucker, K.A., Oliveto, A., Prince, C., Sinha, R., Po-tenza,M.N., Skudlarski, P., andWexler, B.E. (2006). Cue-induced brain activitychanges and relapse in cocaine-dependent patients. Neuropsychopharma-cology 31, 644–650.

Kumari, V., Peters, E.R., Fannon, D., Antonova, E., Premkumar, P., Anilkumar,A.P., Williams, S.C., and Kuipers, E. (2009). Dorsolateral prefrontal cortex ac-tivity predicts responsiveness to cognitive-behavioral therapy in schizo-phrenia. Biol. Psychiatry 66, 594–602.

Kumari, V., Antonova, E., Fannon, D., Peters, E.R., Ffytche, D.H., Premkumar,P., Raveendran, V., Andrew, C., Johns, L.C., McGuire, P.A., et al. (2010).Beyond dopamine: functional MRI predictors of responsiveness to cognitivebehaviour therapy for psychosis. Front. Behav. Neurosci. 4, 4.

Kumari, V., Fannon, D., Peters, E.R., Ffytche, D.H., Sumich, A.L., Premkumar,P., Anilkumar, A.P., Andrew, C., Phillips, M.L., Williams, S.C., and Kuipers, E.(2011). Neural changes following cognitive behaviour therapy for psychosis: alongitudinal study. Brain 134, 2396–2407.

Levy, I., Lazzaro, S.C., Rutledge, R.B., and Glimcher, P.W. (2011). Choice fromnon-choice: predicting consumer preferences from blood oxygenation level-dependent signals obtained during passive viewing. J. Neurosci. 31, 118–125.

Lopez, R.B., Hofmann, W., Wagner, D.D., Kelley, W.M., and Heatherton, T.F.(2014). Neural predictors of giving in to temptation in daily life. Psychol. Sci.25, 1337–1344.

Lopez-Barroso, D., Catani, M., Ripolles, P., Dell’Acqua, F., Rodrıguez-For-nells, A., and de Diego-Balaguer, R. (2013). Word learning is mediated bythe left arcuate fasciculus. Proc. Natl. Acad. Sci. USA 110, 13168–13173.

Mahmood, O.M., Goldenberg, D., Thayer, R., Migliorini, R., Simmons, A.N.,and Tapert, S.F. (2013). Adolescents’ fMRI activation to a response inhibitiontask predicts future substance use. Addict. Behav. 38, 1435–1441.

Maisto, S.A., and Connors, G.J. (2006). Relapse in the addictive behaviors:integration and future directions. Clin. Psychol. Rev. 26, 229–231.

Mather, M., Cacioppo, J.T., and Kanwisher, N. (2013). Introduction to the spe-cial section 20 years of fMRI—What has it done for understanding cognition?Perspect. Psychol. Sci. 8, 41–43.

Maurer, U., Bucher, K., Brem, S., Benz, R., Kranz, F., Schulz, E., van der Mark,S., Steinhausen, H.C., and Brandeis, D. (2009). Neurophysiology in preschoolimproves behavioral prediction of reading ability throughout primary school.Biol. Psychiatry 66, 341–348.

Mayberg, H.S., Brannan, S.K., Mahurin, R.K., Jerabek, P.A., Brickman, J.S.,Tekell, J.L., Silva, J.A., McGinnis, S., Glass, T.G., Martin, C.C., and Fox, P.T.(1997). Cingulate function in depression: a potential predictor of treatmentresponse. Neuroreport 8, 1057–1061.

Mayberg, H.S., Lozano, A.M., Voon, V., McNeely, H.E., Seminowicz, D., Ha-mani, C., Schwalb, J.M., and Kennedy, S.H. (2005). Deep brain stimulationfor treatment-resistant depression. Neuron 45, 651–660.

Neuron

Review

McClure, E.B., Adler, A., Monk, C.S., Cameron, J., Smith, S., Nelson, E.E., Lei-benluft, E., Ernst, M., and Pine, D.S. (2007). fMRI predictors of treatmentoutcome in pediatric anxiety disorders. Psychopharmacology (Berl.) 191,97–105.

McGrath, C.L., Kelley, M.E., Dunlop, B.W., Holtzheimer, P.E., III, Craighead,W.E., and Mayberg, H.S. (2013a). Pretreatment brain states identify likely fail-ures to standard treatments for depression. Biol. Psychiatry 76, 527–535.

McGrath, C.L., Kelley, M.E., Holtzheimer, P.E., Dunlop, B.W., Craighead,W.E.,Franco, A.R., Craddock, R.C., and Mayberg, H.S. (2013b). Toward a neuroi-maging treatment selection biomarker for major depressive disorder. JAMAPsychiatry 70, 821–829.

McNorgan, C., Alvarez, A., Bhullar, A., Gayda, J., and Booth, J.R. (2011). Pre-diction of reading skill several years later depends on age and brain region: im-plications for developmental models of reading. J. Neurosci. 31, 9641–9648.

Miller, W.R. (1996). What is a relapse? Fifty ways to leave the wagon. Addiction91 (Suppl ), S15–S27.

Milstein, R.M., Wilkinson, L., Burrow, G.N., and Kessen, W. (1981). Admissiondecisions and performance during medical school. J. Med. Educ. 56, 77–82.

Mitte, K. (2005). Meta-analysis of cognitive-behavioral treatments for general-ized anxiety disorder: a comparisonwith pharmacotherapy. Psychol. Bull. 131,785–795.

Molfese, D.L. (2000). Predicting dyslexia at 8 years of age using neonatal brainresponses. Brain Lang. 72, 238–245.

Monahan, J. (1981). Predicting Violent Behavior: An Assessment of ClinicalTechniques. (Sage Publications).

Murdaugh, D.L., Cox, J.E., Cook, E.W., 3rd, andWeller, R.E. (2012). fMRI reac-tivity to high-calorie food pictures predicts short- and long-term outcome in aweight-loss program. Neuroimage 59, 2709–2721.

Naqvi, N.H., Rudrauf, D., Damasio, H., and Bechara, A. (2007). Damage to theinsula disrupts addiction to cigarette smoking. Science 315, 531–534.

Noel, X., Sferrazza, R., Van Der Linden, M., Paternot, J., Verhas, M., Hanak, C.,Pelc, I., and Verbanck, P. (2002). Contribution of frontal cerebral blood flowmeasured by (99m)Tc-Bicisate spect and executive function deficits to pre-dicting treatment outcome in alcohol-dependent patients. Alcohol Alcohol.37, 347–354.

Norman, A.L., Pulido, C., Squeglia, L.M., Spadoni, A.D., Paulus, M.P., and Ta-pert, S.F. (2011). Neural activation during inhibition predicts initiation of sub-stance use in adolescence. Drug Alcohol Depend. 119, 216–223.

Otto, M.W., Pollack, M.H., Gould, R.A., Worthington, J.J., 3rd, McArdle, E.T.,Rosenbaum, J.F., and Heimberg, R.G. (2000). A comparison of the efficacy ofclonazepam and cognitive-behavioral group therapy for the treatment of socialphobia. J. Anxiety Disord. 14, 345–358.

Paulus, M.P., Tapert, S.F., and Schuckit, M.A. (2005). Neural activation pat-terns of methamphetamine-dependent subjects during decision making pre-dict relapse. Arch. Gen. Psychiatry 62, 761–768.

Pennington, B.F., and Gilger, J.W. (1996). How is dyslexia transmitted? InDevelopmental Dyslexia: Neural, Cognitive, and Genetic Mechanisms, C.H.Chase, G.D. Rosen, and G.F. Sherman, eds. (York Press), pp. 41–61.

Pereira, F., Mitchell, T., and Botvinick, M. (2009). Machine learning classifiersand fMRI: a tutorial overview. Neuroimage 45 (1, Suppl), S199–S209.

Pizzagalli, D.A. (2011). Frontocingulate dysfunction in depression: toward bio-markers of treatment response. Neuropsychopharmacology 36, 183–206.

Pollack, M.H., Meoni, P., Otto, M.W., Simon, N., and Hackett, D. (2003). Pre-dictors of outcome following venlafaxine extended-release treatment of DSM-IV generalized anxiety disorder: a pooled analysis of short- and long-termstudies. J. Clin. Psychopharmacol. 23, 250–259.

Premkumar, P., Fannon, D., Kuipers, E., Peters, E.R., Anilkumar, A.P., Sim-mons, A., and Kumari, V. (2009). Structural magnetic resonance imagingpredictors of responsiveness to cognitive behaviour therapy in psychosis.Schizophr. Res. 115, 146–155.

Prisciandaro, J.J., Myrick, H., Henderson, S., McRae-Clark, A.L., and Brady,K.T. (2013). Prospective associations between brain activation to cocaineand no-go cues and cocaine relapse. Drug Alcohol Depend. 131, 44–49.

Qi, Z., Han, M., Garel, K., Chen, E.S., and Gabrieli, J.D.E. (2014). White-matterstructure in the right hemisphere predicts Mandarin Chinese learning success.J. Neurolinguist. Published online September 2, 2014. http://dx.doi.org/10.1016/j.jneuroling.2014.08.004.

Raghubar, K.P., Barnes, M.A., and Hecht, S.A. (2010). Working memory andmathematics: A review of developmental, individual difference, and cognitiveapproaches. Learn. Individ. Differ. 20, 110–122.

Rando, K., Hong, K.I., Bhagwagar, Z., Li, C.S., Bergquist, K., Guarnaccia, J.,and Sinha, R. (2011). Association of frontal and posterior cortical gray mattervolume with time to alcohol relapse: a prospective study. Am. J. Psychiatry168, 183–192.

Rao, R., Fung, G., and Rosales, R. (2008). On the dangers of cross-validation.An experimental evaluation. In Proceedings of the 2008 SIAM InternationalConference on Data Mining, pp. 588–596.

Salomons, T.V., Dunlop, K., Kennedy, S.H., Flint, A., Geraci, J., Giacobbe, P.,and Downar, J. (2014). Resting-state cortico-thalamic-striatal connectivitypredicts response to dorsomedial prefrontal rTMS in major depressive disor-der. Neuropsychopharmacology 39, 488–498.

Scoville, W.B., and Milner, B. (1957). Loss of recent memory after bilateral hip-pocampal lesions. J. Neurol. Neurosurg. Psychiatry 20, 11–21.

Shaywitz, S.E. (1998). Dyslexia. N. Engl. J. Med. 338, 307–312.

Shehzad, Z., Kelly, A.M., Reiss, P.T., Gee, D.G., Gotimer, K., Uddin, L.Q., Lee,S.H., Margulies, D.S., Roy, A.K., Biswal, B.B., et al. (2009). The resting brain:unconstrained yet reliable. Cereb. Cortex 19, 2209–2229.

Siegle, G.J., Carter, C.S., and Thase, M.E. (2006). Use of FMRI to predict re-covery from unipolar depression with cognitive behavior therapy. Am. J. Psy-chiatry 163, 735–738.

Siegle, G.J., Thompson, W.K., Collier, A., Berman, S.R., Feldmiller, J., Thase,M.E., and Friedman, E.S. (2012). Toward clinically useful neuroimaging indepression treatment: prognostic utility of subgenual cingulate activity fordetermining depression outcome in cognitive therapy across studies, scan-ners, and patient characteristics. Arch. Gen. Psychiatry 69, 913–924.

Simpson, H.B., Foa, E.B., Liebowitz, M.R., Huppert, J.D., Cahill, S., Maher,M.J., McLean, C.P., Bender, J., Jr., Marcus, S.M., Williams, M.T., et al.(2013). Cognitive-behavioral therapy vs risperidone for augmenting serotoninreuptake inhibitors in obsessive-compulsive disorder: a randomized clinicaltrial. JAMA Psychiatry 70, 1190–1199.

Sorg, S.F., Taylor, M.J., Alhassoon, O.M., Gongvatana, A., Theilmann, R.J.,Frank, L.R., and Grant, I. (2012). Frontal white matter integrity predictors ofadult alcohol treatment outcome. Biol. Psychiatry 71, 262–268.

Souery, D., Papakostas, G.I., and Trivedi, M.H. (2006). Treatment-resistantdepression. J. Clin. Psychiatry 67 (Suppl 6 ), 16–22.

Stice, E., Yokum, S., Bohon, C., Marti, N., and Smolen, A. (2010). Reward cir-cuitry responsivity to food predicts future increases in body mass: moderatingeffects of DRD2 and DRD4. Neuroimage 50, 1618–1625.

Stice, E., Yokum, S., and Burger, K.S. (2013). Elevated reward region respon-sivity predicts future substance use onset but not overweight/obesity onset.Biol. Psychiatry 73, 869–876.

Supekar, K., Swigart, A.G., Tenison, C., Jolles, D.D., Rosenberg-Lee, M.,Fuchs, L., and Menon, V. (2013). Neural predictors of individual differencesin response to math tutoring in primary-grade school children. Proc. Natl.Acad. Sci. USA 110, 8230–8235.

Tan, L.H., Chen, L., Yip, V., Chan, A.H., Yang, J., Gao, J.-H., and Siok, W.T.(2011). Activity levels in the left hemisphere caudate-fusiform circuit predicthow well a second language will be learned. Proc. Natl. Acad. Sci. USA 108,2540–2544.

Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. J. R.Stat. Soc. Series B Stat. Methodol. 58, 267–288.

Neuron 85, January 7, 2015 ª2015 Elsevier Inc. 25

Neuron

Review

Udo, T., Clifford, P.R., Davis, C.M., and Maisto, S.A. (2009). Alcohol use postAUD treatment initiation as a predictor of later functioning. Am. J. Drug AlcoholAbuse 35, 128–132.

Ullman, H., Almeida, R., and Klingberg, T. (2014). Structural maturation andbrain activity predict future working memory capacity during childhood devel-opment. J. Neurosci. 34, 1592–1598.

Vapnik, V.N. (1999). An overview of statistical learning theory. IEEE Trans. Neu-ral Netw. 10, 988–999.

Ventura-Campos, N., Sanjuan, A., Gonzalez, J., Palomar-Garcıa, M.A., Rodrı-guez-Pujadas, A., Sebastian-Galles, N., Deco, G., and Avila, C. (2013). Spon-taneous brain activity predicts learning ability of foreign sounds. J. Neurosci.33, 9295–9305.

Vollmar, C., O’Muircheartaigh, J., Barker, G.J., Symms, M.R., Thompson, P.,Kumari, V., Duncan, J.S., Richardson, M.P., and Koepp, M.J. (2010). Identical,but not the same: intra-site and inter-site reproducibility of fractional anisot-ropy measures on two 3.0T scanners. Neuroimage 51, 1384–1394.

Wager, T.D., Atlas, L.Y., Leotti, L.A., and Rilling, J.K. (2011). Predicting individ-ual differences in placebo analgesia: contributions of brain activity duringanticipation and pain experience. J. Neurosci. 31, 439–452.

Walkup, J.T., Albano, A.M., Piacentini, J., Birmaher, B., Compton, S.N., Sher-rill, J.T., Ginsburg, G.S., Rynn, M.A., McCracken, J., Waslick, B., et al. (2008).Cognitive behavioral therapy, sertraline, or a combination in childhood anxiety.N. Engl. J. Med. 359, 2753–2766.

Walsh, B.T., Seidman, S.N., Sysko, R., and Gould, M. (2002). Placeboresponse in studies of major depression: variable, substantial, and growing.JAMA 287, 1840–1847.

Wechsler, H., Lee, J.E., Nelson, T.F., and Kuo, M.; Findings from the HarvardSchool of Public Health College Alcohol Study (2002). Underage college stu-dents’ drinking behavior, access to alcohol, and the influence of deterrencepolicies. J. Am. Coll. Health 50, 223–236.

Wender, P.H. (1998). Pharmacotherapy of attention-deficit/hyperactivity dis-order in adults. J. Clin. Psychiatry 59 (Suppl 7 ), 76–79.

Whalen, P.J., Johnstone, T., Somerville, L.H., Nitschke, J.B., Polis, S., Alex-ander, A.L., Davidson, R.J., and Kalin, N.H. (2008). A functional magnetic reso-nance imaging predictor of treatment response to venlafaxine in generalizedanxiety disorder. Biol. Psychiatry 63, 858–863.

26 Neuron 85, January 7, 2015 ª2015 Elsevier Inc.

Whelan, R., and Garavan, H. (2014). When optimism hurts: inflated predictionsin psychiatric neuroimaging. Biol. Psychiatry 75, 746–748.

Whelan, R., Watts, R., Orr, C.A., Althoff, R.R., Artiges, E., Banaschewski, T.,Barker, G.J., Bokde, A.L.W., Buchel, C., Carvalho, F.M., et al.; IMAGEN Con-sortium (2014). Neuropsychosocial profiles of current and future adolescentalcohol misusers. Nature 512, 185–189.

Wolff, J.J., Gu, H., Gerig, G., Elison, J.T., Styner, M., Gouttard, S., Botteron,K.N., Dager, S.R., Dawson, G., Estes, A.M., et al.; IBIS Network (2012). Differ-ences in white matter fiber tract development present from 6 to 24 months ininfants with autism. Am. J. Psychiatry 169, 589–600.

Wonderlick, J.S., Ziegler, D.A., Hosseini-Varnamkhasti, P., Locascio, J.J.,Bakkour, A., van der Kouwe, A., Triantafyllou, C., Corkin, S., and Dickerson,B.C. (2009). Reliability of MRI-derived cortical and subcortical morphometricmeasures: effects of pulse sequence, voxel geometry, and parallel imaging.Neuroimage 44, 1324–1333.

World Health Organization (2008). WHO report on the global tobaccoepidemic. http://www.who.int/tobacco/mpower/mpower_report_tobacco_crisis_2008.pdf

Wykes, T., Steel, C., Everitt, B., and Tarrier, N. (2008). Cognitive behavior ther-apy for schizophrenia: effect sizes, clinical models, and methodological rigor.Schizophr. Bull. 34, 523–537.

Yang, M., Wong, S.C., and Coid, J. (2010). The efficacy of violence prediction:a meta-analytic comparison of nine risk assessment tools. Psychol. Bull. 136,740–767.

Yokum, S., Ng, J., and Stice, E. (2011). Attentional bias to food images asso-ciated with elevated weight and future weight gain: an fMRI study. Obesity(Silver Spring) 19, 1775–1783.

Yoo, J.J., Hinds, O., Ofen, N., Thompson, T.W., Whitfield-Gabrieli, S., Trianta-fyllou, C., and Gabrieli, J.D. (2012). When the brain is prepared to learn:enhancing human learning using real-time fMRI. Neuroimage 59, 846–852.

Zatorre, R.J. (2013). Predispositions and plasticity in music and speechlearning: neural correlates and implications. Science 342, 585–589.

Zatorre, R.J., Delhommeau, K., and Zarate, J.M. (2012). Modulation of auditorycortex response to pitch variation following training with microtonal melodies.Front. Psychol. 3, 544.